The surface water trap: risk for banks and insurers - Part 1

Part 1: What surface water does to collateral

It was 21 August 2020. I was sitting at the head office of the bank I worked for in Tønsberg, looking out over a pedestrian underpass. The rain was so torrential that a Norwegian record for rainfall in a single hour was set in the neighbourhood, and the underpass flooded. Out of the corner of my eye, I saw someone walk into it anyway. Five minutes later, three grown men in high-vis vests were standing chest-deep in water down there. That was when I realised the man I had just seen walk in could have drowned, and that this was probably what they were checking.

From peripheral problem to systemic risk

Back then, Finanstilsynet, the Financial Supervisory Authority of Norway, considered transition risk the greatest short-term threat to financial stability, and believed surface water risk like this lay further in the future. But average temperatures in Norway have risen, wind and rainfall patterns are changing, and this is hitting buildings and infrastructure that are rarely adapted to the climate. The result has been rising insurance payouts: around half of annual weather-related payouts to households relate to damage not covered by the natural perils scheme, with surface water the most important cause, and payouts rose sharply from 2008 to 2024, especially after Storm Hans in 2023. From 2016 to 2024, the price of home insurance rose by more than 60 percent in fixed 2024 kroner. From 2024 to 2025, the statutory natural perils premium rose by 23 percent.

Municipal charges grew by more than 115 percent from 2008 to 2024, partly because of wastewater treatment, but also because of a changing climate. Under cost-recovery rules, water and wastewater services must be financed krone for krone by what they cost the municipality, so when climate change demands larger pipes and more detention basins, the cost lands straight on our bills. A survey by Huseierne, the Norwegian homeowners' association, of the 120 largest municipalities shows a 60 percent increase in water charges and 57 percent in wastewater charges from 2023 to 2027. That cost growth eats into the debt-servicing capacity of the bank's borrowers.

Altogether, the cost of home insurance and municipal charges for an average detached house has almost doubled in just under ten years, from around 15 000 to nearly 30 000 kroner a year, measured in fixed 2024 kroner. And the number of days with very heavy rainfall is likely to keep increasing across almost the entire country. Home ownership is very common in Norway and household debt is high, so rising housing costs weaken people's ability to handle unexpected expenses. Risk zones for natural perils, where banks and borrowers are less exposed thanks to the natural perils scheme, are well mapped. Ironically, surface water, a very widespread uninsured risk, is poorly mapped by the public sector. But the insurers, who carry the risk themselves, have mapped it. Now they are competing to price it correctly. Homes and commercial buildings with a claims history or in risk zones may therefore face higher premiums or demands for remedial work. Over time, this may mean fewer people can afford insurance. Let's hope they tell the bank when they cancel it.

What the numbers say about property values and surface water risk

Staff at Norges Bank have published a report linking surface water risk to property values and debt-servicing capacity, using data from our partner 7Analytics. It shows that 5.4 percent of Norwegian homes, and 5.6 percent of housing value, have high surface water risk. 7Analytics, together with Housing Lab at OsloMet, has also found that moving from 0 to 1 on the flood risk scale is associated with a price around 5.6 percent lower. The finding concerns flood risk. However, floods and surface water often affect the same areas, so it suggests how much prices could fall where the risk is surface water too. (The similarity to the 5.6 percent above is a coincidence: one is a price difference, the other a share of housing value.) This can reduce collateral values in exposed areas.

The Norges Bank report found that the share of lending by the 18 largest banks to detached and small houses with high surface water risk ranges from under 3 to around 10 percent. The maximum would probably have been higher had smaller, geographically concentrated banks been included in the analysis: Vestfold, Møre og Romsdal, Nordland, and Troms and Finnmark have the highest share of housing value at high risk, while Akershus, Rogaland and Trøndelag have the largest absolute housing values in the risk zones. Commercial property may be built with greater climate risk than housing, so the spread is unlikely to be any narrower there. In several counties, housing values in the risk zones are lower than the number of homes would suggest, meaning the most exposed homes are low priced. The price level determines whether climate adaptation pays off. Møre og Romsdal is among the counties with the highest share, and a remote example from there illustrates the mechanism. On Smøla, the average home sells for around 1.9 million kroner. The price window in the local housing market is then probably so narrow that if you invest half a million in climate adaptation, you are unlikely to get it back in a correspondingly higher sale price.

Surface water is not only one thing

Buildings with fire insurance are automatically covered for natural perils, and the Natural Perils Insurance Act covers, among other things, floods, meaning water from watercourses that overflow their banks. The premium is set by law and is the same for all buildings, however exposed they are. But much of what floods Norwegian basements is not flooding in this sense. Floods and surface water often occur at the same time, in the same place, but they are insured differently. The distinction decides who pays: the natural perils scheme, which everyone contributes to through the statutory premium, or the individual owner through their own insurance and deductible.

The natural perils scheme does not cover surface water, meaning rainfall that has no time to soak into the ground or drain away through the sewer network. That is covered, at best, by the water damage cover in the ordinary buildings insurance. The assumption that "insurance will take care of it" holds no better for surface water than it did for fire in our previous article series on Krokstadelva, and partly for the same reasons:

The gaps insurance doesn't plug for the bank

Reduced settlements after repeated surface water damage are not unusual. Insurance usually covers only restoration, not prevention, so the underlying exposure doesn't change: once a property has had water in the basement, the insurer often expects the owner to protect the building with a backflow valve, raised drains or regraded terrain. If those measures are missing next time, the settlement can be reduced or refused, and cover can shrink significantly from one event to the next without the bank knowing. That is a double risk: higher default risk for the customer and a loss in collateral value (more in part 3).

Responsibility for surface water events is split between the municipality, the insurer and the homeowner, and none of them has the full picture. A bank assessing collateral with surface water risk is in practice assessing three factors it doesn't control: the investment pace of a municipality that has neither the authority nor the budget to manage surface water on private land, the owner's compliance with requirements, and how the tug-of-war between insurer and municipality plays out if disaster strikes. The warranty period in typical construction contracts (five years) is shorter than the time horizon of the damage, so the contractor doesn't stand behind the building's condition when the next incident occurs either. If no party guarantees the residual value, the bank can quickly be left holding the uninsured risk.

The bank's concentration risk follows the catchment area. Surface water damage rarely hits one address at a time. It hits the low point of a catchment, however many buildings are located there and whoever owns them. This is not an individual risk, but correlated exposure for banks with geographically concentrated portfolios. Severity can still vary from building to building on the same street, depending on whether it is exposed, old or new, whether the owner has taken measures, or whether it has a basement.

A loan for a home or a commercial building often runs for 25 years or more, even if it is typically refinanced along the way, while the insurance policy is renewed every year. Unlike the statutory natural perils premium, the water damage element of buildings insurance can be priced by risk, so customers whose insurer discovers rising surface water risk during the life of the loan may face substantially higher premiums. That probably increases the risk of the customer defaulting on their loan.

Inadequate insurance cover can increase the bank's loss given default, and this can happen in several ways. Cover can be missing entirely: around one in ten households lack contents insurance, some also buildings insurance, and if one or both are missing, servicing the mortgage becomes difficult. Cover can also be too narrow. Nesbyen during Storm Hans shows how the distinction between flood and surface water plays out. A building there could in principle have been fully insured against the flood during Hans, but underinsured against water coming through the drainage system if it only had standard cover. The same storm then caused two types of damage, each with its own payer. The sequence of causes can also affect how much of a surface water claim the owner gets covered. Did the water come from outside as surface runoff, or from below as sewer backflow? The source also determines whether the damage counts as a natural peril: a "wild stream" is formally a watercourse and therefore regulated by law, while surface runoff from a newly asphalted area with too few drains depends on policy terms. Full-value insurance is meant to cover rebuilding, but older buildings with a maintenance backlog don't always get full value, and commercial buildings and holiday homes often have a fixed sum insured that is eroded by inflation.

If a commercial building has to relocate because of surface water damage, the insurance must specifically cover that. If a home is hit by surface water and becomes subject to a building and subdivision ban, the house is compensated, but not the fall in the plot's value. A restricted plot still has value, but clearly less than one that is ready to build on. The difference ends up with the borrower, who is still servicing the loan on the whole property, and therefore with the bank. Plot values can also fall for neighbours without damage. In central Nesbyen, flood-affected residents are still living away from home three years after the event, because a municipal building and subdivision ban pending a flood barrier study is a precondition for a full insurance settlement. The example concerns flooding, but the mechanism is the same for surface water: the municipality can restrict areas while it investigates measures, and the loss in value falls outside the insurance.

What this means for credit work

Today, most Norwegian banks rely on map lookups against NVE's flood zone maps when granting loans, if they use anything at all (some use them only for homes, even though commercial buildings are often larger individual exposures). The flood zone maps don't show surface water. That is a paradox: banks check the risk that is largely already covered by the natural perils scheme, but not the one that largely isn't. The insurers covering the buildings the banks hold as collateral, by contrast, work with better and more granular data on where surface water can actually flow during a cloudburst. The banks' analysis also doesn't show whether the policy covers the damage in full, whether the last claim triggered the reduction clause, or whether the catchment contains eleven other properties pledged to the same bank. That weakens the precision of estimates for probability of default and loss given default. Correlated surface water risk is also relevant for capital requirement assessments.

Telescope now gives banks and insurers data on surface water risk for each individual building at an address, weighted against historical claims statistics, to make the differences measurable. That is also the starting point for part 2: what regulation requires banks to be able to document about surface water risk, and what it can cost not to be able to.

This is part 1 of three on surface water and credit risk. Part 2 covers the documentation requirements in the capital requirements framework. Part 3 is a worked example putting numbers on what repeated damage can do to loss given default.